Wanted: A Silver Bullet MLOps Solution for Enterprise

The pharmaceutical research and early development (pRED) business unit in Roche faces increasing demand for operationalizing Machine Learning models. This talk shares key learnings and recommendations from their 4-year MLOps journey, discussing how they sustainably served diverse users and use cases despite a fragmented and volatile tool landscape. They also share perspectives on the future of MLOps in their research organization, highlighting challenges and lessons learned in building MLOps solutions, including persona modeling, an API-first approach, and the 'less is more' principle.